简介:WeproposeanefficientandrobustalgorithmtosolvethesteadyEulerequa-tionsonunstructuredgrids.ThenewalgorithmisaNewton-iterationmethodinwhicheachiterationstepisalinearmultigridmethodusingblocklower-uppersymmetricGauss-Seidel(LU-SGS)iterationasitssmootherToregularizetheJacobianmatrixofNewton-iteration,weadoptedalocalresidualdependentregularizationasthereplace-mentofthestandardtime-steppingrelaxationtechniquebasedonthelocalCFLnumberTheproposedmethodcanbeextendedtohighorderapproximationsandthreespatialdimensionsinanatureway.Thesolverwastestedonasequenceofbenchmarkprob-lemsonbothquasi-uniformandlocaladaptivemeshes.Thenumericalresultsillustratedtheefficiencyandrobustnessofouralgorithm.